Subnational probabilistic projections of fertility: rethinking from Latin America

Abstract

General trends in fertility, mortality, and migration can be discerned and projected into the future with reasonable results, however, there is considerable uncertainty attached to each specific trend from a particular country or region. Then, subnational projections then represent a special chapter within the demographic projections. After introducing subnational projections and its close relationship with the fertility, this work proposes as the main objective to project fertility rates at the subnational level for Argentina using a probabilistic method; the bayesian hierarchical model (BHM), and then compare the results with the point estimates of deterministic projections published by the national organism of statistics. Forecasts were obtained from two models, one including all the countries available in World Population Prospects (WPP) and a second model based only in a subgroup of countries; Argentina, Colombia, Chile, Cuba and Uruguay, both based on data from 1980 to 2010. This set of countries where selected using the transitional to identify a subset of countries with similar patterns among them, that includes Argentina, but show different patterns compared to the rest region. Time period is selected to include years for which these countries reach an adequate data quality. BHM for subnational projections is an extremely useful and flexible method. It presents many advantages over the classical methods. Bayesian framework is a powerful scheme to generate national and subnational projections for mortality, fertility and finally population. This paper reinforces the use of probabilistic models, moreover BHM, that respects data not forcing it to get caught into a mathematical assumption.